KR20160136442A - Computer generated natural language outputs in question answer systems - Google Patents

Abstract

Methods for generating natural language output, computer systems, and computer storage media are provided. The set of triples can be used to map voice queries and responses to a sentence structure that can be used as an output response to a voice query. The sentence structure is suitable only for a predetermined triple set. One or more constraints may be associated with a set of triples to ensure that the sentence structure is applied only in the correct context. To become a valid sentence structure, each constraint associated with the sentence structure must be satisfied. If each constraint is satisfied, the sentence structure is valid and can be used as a form for the output response. If each constraint is not satisfied, the additional sentence structure associated with the triple set can be evaluated until the valid sentence structure is satisfied. If the sentence structure is invalid, no output is generated.

Description

[0001] COMPUTER GENERATED NATURAL LANGUAGE OUTPUTS IN QUESTION ANSWER SYSTEMS [0002]

The present invention relates to a computer generated natural language output in a query response system.

Natural language generation is generally performed using a set of triples from a knowledge base. Triples are compiled into valid natural language statements. It is increasingly common for a user to output a natural language sentence in response to a verbal query to the device. Sometimes, in response to a query that is spontaneously expressed, the natural language sentence response that comes back is inaccurate, robotic, and unnatural.

This summary is provided to introduce a selection of concepts in a simplified form that are more fully described in the detailed description which follows. This summary is not intended to identify key features or essential features of the claimed subject matter and is not intended to be used as an aid in determining the scope of the claimed subject matter.

Embodiments of the present invention are particularly directed to systems, methods and computer storage media for generating natural language output. As noted, the present invention seeks to produce a natural language output that is accurate (or valid) natural sound in that it is a well-flowing sentence when the output is spoken. Knowledge base triples can be used to represent queries in a machine-readable language that can be translated into natural language output. A triple may be associated with one or more sentence structures suitable for use with a particular triple. The sentence structure used herein refers to an exemplary context free sentence format that typically includes one or more substitutable variables. The triple and / or sentence structure may also be associated with one or more constraints. The constraints used here generally refer to rules that restrict the types of values that can be substituted for variables. The sentence structure can be used as an output response when the constraint associated with the sentence structure is satisfied, as will be described later.

Thus, in one embodiment, the invention relates to one or more computer storage media embodied in computer-executable instructions for performing a method of generating a natural language output when executed by one or more computing devices. The method includes receiving a query from a user; Identifying a response to the query; Mapping the response to structured data from a knowledge base; Identifying a sentence structure associated with the structured data; Confirming that at least one constraint associated with the sentence structure is satisfied; And delivering an output response for the query in the form of a statement when each of the one or more constraints is satisfied.

In another embodiment, the invention relates to a computer system for generating natural language output. A system includes a computing device associated with a natural language engine having one or more processors and one or more computer storage media; Wherein the natural language engine identifies a response to a query, maps the response to structured data from a knowledge base, identifies a sentence structure associated with the structured data, Identifies one or more constraints associated with the sentence structure, and conveys the output response for the query in the form of a sentence.

In another embodiment, the invention relates to a computerized method of generating a natural language output. The method includes receiving a query from a user; Identifying a response to the speech input query; Mapping the response to a set of triples; At least one rule associated with the set of triples, the at least one rule comprising a context free grammar sentence structure associated with a set of triples and at least one constraint associated with the context free grammar sentence; Determining if at least one constraint is satisfied; And delivering a speech output response to the speech input query to the sentence when determining that at least one constraint associated with the context free grammar sentence is satisfied.

The present invention is described in detail below with reference to the accompanying drawings.
1 is a block diagram of an exemplary computing environment suitable for use in implementing an embodiment of the invention.
2 is a block diagram of an exemplary system for generating a natural language output suitable for use in implementing an embodiment of the present invention.
3 is a flow diagram of an exemplary method for generating a natural language output, in accordance with an embodiment of the present invention.
4 is a flow diagram of an exemplary method for generating a natural language output, in accordance with an embodiment of the present invention.
5 is a flow diagram of an exemplary method for generating a natural language output, in accordance with an embodiment of the present invention.

The subject matter of the present invention is specifically described herein to meet legal requirements. However, the description itself is not limited to limiting the scope of the present patent. Rather, the inventors contemplate that the claimed subject matter may also be practiced in other ways, including other steps or combinations of steps similar to those described in this specification, along with other current or future techniques. Furthermore, although the terms "step" and / or "block" may be used herein to encompass other method elements that are used, they are used interchangeably unless the order of the individual steps is explicitly stated, It should not be interpreted as implying any particular order between the various steps disclosed herein.

Various aspects of the techniques described herein generally relate to, among other things, systems, methods, and computer storage media that generate natural language output. The present invention relates to generating a natural language output that is accurate (or valid) and natural sound in that the output is a well-flowing sentence when spoken. Knowledge base triples can be used to represent queries in a machine-readable language. A triple can be associated with a sentence structure suitable for use with a particular triple. The sentence structure used herein refers to an exemplary context free sentence format that typically includes one or more substitutable variables. The triple and / or sentence structure may also be associated with one or more constraints. The constraints used here generally refer to rules that restrict the types of values that can be substituted for variables. The sentence structure can be used as an output response when the constraint associated with the sentence structure is satisfied.

An overview of an embodiment of the present invention is briefly described and an exemplary operating environment in which embodiments of the present invention may be implemented is described below to provide a general context for various aspects of the present invention. Referring generally to the drawings and particularly to FIG. 1, an exemplary operating environment for implementing an embodiment of the present invention is shown and generally designated as computing device 100. However, the computing device 100 is an example of a suitable computing environment and is not intended to limit the scope of use or functionality of the embodiments of the present invention. The computing device 100 should not be construed as having any dependency or requirement relating to any one or combination of the illustrated components.

Embodiments of the invention may be implemented as computer code or machine-executable instructions, including computer-usable or computer-executable instructions, such as program modules, executed by a computer, such as a personal digital assistant, smartphone, tablet PC or other handheld device, As shown in FIG. Generally, program modules, including routines, programs, objects, components, data structures, etc., refer to code that performs a particular task or implements a particular abstract data type. Embodiments of the invention may be practiced with a variety of system configurations including handheld devices, consumer electronics, general purpose computers, special purpose computing devices, and the like. Embodiments of the invention may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communications network. In a distributed computing environment, program modules may be located on local and remote computer storage media including memory storage devices.

1, computing device 100 includes a memory 112, one or more processors 114, one or more presentation components 116, one or more input / output (I / O) ports 118, And includes a bus 110 that directly or indirectly couples one or more I / O components 120 and an exemplary power supply 122. Bus 110 represents one or more buses (e.g., an address bus, a data bus, or a combination thereof). Although the various blocks of FIG. 1 are shown as lines for clarity, in reality these blocks do not necessarily have to be practical and represent logical components. For example, a presentation component such as a display device that is an I / O component can be considered. The processor also has a memory. The inventors recognize that these are characteristic of the technology and that the drawing of Figure 1 is merely an example of an exemplary computing device that may be used in conjunction with one or more embodiments of the present invention. All of the categories such as "workstation", "server", "laptop", "handheld device", etc. are all within the scope of FIG.

Computing device 100 typically includes a variety of computer readable media. Computer readable media can be any available media accessible by computing device 100 and includes both volatile and nonvolatile, removable and non-removable media. The computer readable medium includes a computer storage medium and a communication medium, wherein the computer storage medium excludes the signal itself. Computer storage media includes volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disk (DVD) or other optical disk storage, magnetic cassettes, magnetic tape, Other magnetic storage devices, or any other medium that can be used to store the desired information and which can be accessed by the computing device 100. The computer storage medium does not include the signal itself. Communication media typically embodies computer readable instructions, data structures, program modules or other data in a modulated data signal such as a carrier wave or other transport mechanism and includes any information delivery media. The term "modulated data signal" means a signal having one or more of its characteristics set or a signal modified in such a way that it encodes information in the signal. By way of example, and not limitation, communication media includes wired media such as a wired network or direct-wired connection, and wireless media such as acoustic, RF, infrared, and other wireless media. Any combination of the foregoing should also be included within the scope of computer readable media.

Memory 112 includes computer storage media in the form of volatile and / or nonvolatile memory. The memory may be separate, non-detachable, or a combination thereof. Exemplary hardware devices include semiconductor memory, hard drives, optical disk drives, and the like. Computing device 100 includes one or more processors that read data from various entities such as memory 112 or I / The presentation component 116 presents the data representation to the user or other device. Exemplary presentation components include a display device, a speaker, a printing component, a vibration component, and the like.

The I / O port 118 allows the computing device 100 to be logically coupled to other devices, including the I / O component 120, and some of the I / O components 120 may be embedded. Exemplary I / O components include microphones, joysticks, game pads, satellite dishes, scanners, printers, wireless devices, controllers such as stylus and keyboard and mouse, and a natural user interface (NUI). The NUI handles the air gesture, voice or other physiological input generated by the user. This input may be a request to interact with a search prefix, a search request, an intent suggestion, a request to interact with an entity or sub-entity, or an advertisement, entity or clarification presented by the computing device 100 Tile, action, search history, and the like. Such a request may be sent to a suitable network element for further processing. The NUI may be any combination of speech recognition, touch and stylus recognition, facial recognition, biometrics, perception on the screen or near the screen, air gestures, head and eye tracking, and touch recognition associated with the display of the computing device 100 Lt; / RTI > The computing device 100 may include depth cameras such as a stereoscopic camera system, an infrared camera system, an RGB camera system, and combinations thereof for behavior detection and recognition. In addition, the computing device 100 may include an accelerometer or gyroscope that enables motion detection. The output of the accelerometer or gyroscope is provided on the display of the computing device 100 to produce a momentum augmented reality or virtual reality.

Various aspects of the subject matter described herein may be described in the general context of computer-executable instructions, such as program modules, being executed by a computing device. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. Various aspects of the subject matter described herein may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communications network. In a distributed computing environment, program modules may be located on local and remote computer storage media including memory storage devices.

Although the term "server" is sometimes used herein, the term refers to a search engine, a web browser, a cloud server, a set of one or more processes distributed on one or more computers, one or more standalone storage devices, Combinations, combinations of one or more of the foregoing, and the like.

Referring now to FIG. 2, a block diagram illustrating an exemplary computing system 200 that may utilize an embodiment of the present invention is shown. In general, the computing system 200 represents an environment in which a natural language output is generated.

Computing system 200 generally includes a network 202, a user device 204, a database 206, and a natural language engine 208, although other components are not shown. The network 202 may include, by way of non-limiting example, one or more local area networks (LANs) and / or a wide area network (WANs). These networking environments are commonplace in offices, enterprise-wide computer networks, intranets, and the Internet. Thus, the network 202 is not described further herein.

The database 206 may be any type of data storage device capable of storing data. Thus, the database 206 may be an online repository of data. Databases are common in computer networks and therefore are not described further here.

User device 204 may be any computing device capable of transmitting speech input queries and receiving speech output queries. For example, the computing device 100 of FIG. 1 may be an exemplary user device. In a particular embodiment, user device 204 is a mobile phone.

The natural language engine 208 may be any device capable of implementing the present invention. The natural language engine 208 may be specifically configured to compile the natural language output in response to a query. In one embodiment, the natural language output is a verbal response in the form of a sentence (e.g., a sentence read by the user device 204). In one embodiment, the single sentence is a natural language output. In an alternative embodiment, a plurality of sentences may be output. The natural language engine 208 may include a receiving component 210, an identifying component 212, a mapping component 214, a configuration component 216, a constraint validator 218, and a communication component 220.

The receiving component 210 may be specifically configured to receive one or more queries. The query may be entered by the user via the user device 204, for example. In an embodiment, the query is a speech input query, which means that the query is input by the user rather than through the keyboard. However, the query can be input via the keyboard in some examples.

As mentioned above, the voice input query can be performed by the user. Speech input queries can be spoken in a natural language format. In other words, the speech input query is typically provided in the form of a question such that the user naturally asks the other person, for example. Thus, the meaning of the speech input query can be identified by a parser (not shown). The parsing of search queries for semantic meaning is well known in the art, and any method known to those skilled in the art may be used to identify the meaning of the query. Identifying the meaning of the query increases the likelihood that the identified response in response to the query will be correct.

The identification component 212 may be specifically configured to identify a response to a voice input query. The identification component 212 may use the meaning of the speech input query (identified by the parser) to identify the response. The response may be identified from the database 206. An exemplary speech input query: response pair is:

Voice input query: What is Tom Hanks' job?

Answer: Actor

In this example, the parser can identify the meaning of the speech input query by identifying the job of the source entity (i.e., Tom Hanks).

The mapping component 214 may be specifically configured to map the response to the speech input query to the structured data of the knowledge base. The structured data may be one or more triples. The set of triples used here is the grouping of subject, predicate, and object. Mapping the response to a set of triples translates the natural language speech input query and response into a format that the natural language engine 208 can read.

With respect to the above example, the set of triples may be as follows.

Tom Hanks, profession, actor

The set of triples can also look like this:

Tom Hanks, Current Occupation, Actor

Tom Hanks is the subject, the occupation or current job is the predicate, and the actor is the target or object.

If the response is mapped to a set of triples, the system 200 can identify the sentence structure for the set of triples. The configuration component 216 may be specifically configured to identify one or more sentence structures associated with the set of triples. As described above, the sentence structure is an exemplary context free form with one or more variables that can be replaced by values. In other words, a variable can be replaced by a context. Since each sentence structure is only valid under a given set of circumstances, there is a finite set of valid sentence structures associated with any triple set.

An exemplary sentence structure that can be associated with the triple set listed above (i.e., Tom Hanks, current job, actor) is as follows.

[Source Entity ID] The current job is [Target Entity ID].

In this example, the variables are [Source Entity ID] and [Target Entity ID]. The source entity ID is replaced by Tom Hanks, and the target entity ID is replaced by an actor, yielding the following sentence output:

Output: Tom Hanks' current job is actor.

The sentence structure can be associated with a set of triples and a set of constraints. As mentioned earlier, constraints refer to rules that generally limit the types of values that can be substituted for variables. The constraint is used to ensure that the sentence structure can only be applied to a particular triple set. For example,

[Source Entity ID] The current job is [Target Entity ID].

Can not be applied to the following triple sets.

Tom Hanks, Key, 6 feet

This triple set is clearly associated with a voice input query about Tom Hanks' key. Thus, the sentence structure that details the current job is not relevant. The constraint validation ensures that the valid sentence structure is selected as the speech output response.

The constraint validator 218 may be specifically configured to validate one or more constraints associated with the sentence structure. If more than one statement structure is identified, the constraints associated with the statement structure are evaluated to ensure that the constraints are satisfied. In order to use the sentence structure as a speech output response, each constraint associated with the sentence structure must be satisfied. If each constraint is not satisfied, the other sentence structure can be evaluated until a statement structure in which all constraints are satisfied is identified. For example, consider the following example.

Voice Input Query: How high is Tom Hanks?

Answer: 6 feet

Triple: Tom Hanks, Key, 6 feet

This triple can be associated with the following sentence structure.

[Source Entity ID] is the key whose [Target Entity ID] is pit.

This sentence structure can be particularly related to the following constraints.

Constraint 1: [Source Entity ID] = Person

Constraint 2: [target entity ID] = number

In this example, [Source Entity ID] is Tom Hanks (i.e., person) and [Target Entity ID] is 6 (i.e., a number). Therefore, constraint conditions 1 and 2 are satisfied. In this case, the constraint validator 218 determines that each constraint is satisfied and identifies the statement structure as a valid armed structure. Constraints can limit subject, object, predicate, relationship between subject and object.

Alternatively, if the sentence structure is found not to be a valid sentence structure (e.g., each constraint associated with the sentence structure is not satisfied), the sentence structure will not be selected as a valid sentence structure to use as a speech output response . In that situation, other sentence structures associated with the set of triples are identified and the constraints associated with the sentence structure can be verified. If no valid sentence structure is identified, the output is not delivered.

The communication component 220 may be specifically configured to communicate a voice output response. The voice output response may be a single sentence or a plurality of sentences. For example, in the above example, the voice output response may be "Tom Hanks is 6 feet tall ". The communication component 220 may send this voice output response to the user device 204.

In one embodiment, communication component 220 may communicate a voice output response with a search result page. For example, a voice output response "Tom Hanks can be delivered as text (as well as text as it is six feet tall"), and a search result page is also presented. The search result page may be a source for a voice output response.

In one embodiment, the grading component (not shown) may be configured to rank sentence structures especially when two or more sentence structures are in effect. For example, if a sentence is to be composed in response to a speech input query for a response that includes the place of birth and the date of birth, it may be desirable to first say the date of birth and then say the place of birth. For example, the output response for a voice input query of "Tom Hanks ' s birth and birth date " may be" Tom Hanks was born in Concord City on July 9, 1956 ". This sentence structure can be regarded as a better flow than first listing the place of birth. The grading component may be configured to rank the sentence structure based on various preferences and / or rules established in the system 200 such that a higher grade sentence structure is selected.

In application, a speech input query is received. An exemplary speech input query may be "Who is Tom Hanks' wife?" The answer to this query is "Rita Wilson". This voice input query and response can be mapped to a triple set that looks like "Tom Hanks, marriage, Rita Wilson". This response and the sentence structure associated with the triple can be identified. An exemplary related sentence structure may be as follows.

The [Source Entity ID] is the [Target Entity ID] and the current marital status since [Token].

A set of constraints associated with the sentence structure is then identified. The set of constraints for this special sentence is:

Constraint 1: Source Entity ID = Person

Constraint 2: Target Entity ID = Person

Constraint 3: Marriage between Source Entity ID and Target Entity ID without end date

Constraint 4: token: year

Constraint 3 restricts the sentence structure to the current marriage when there is no end date. In this example, the source entity ID is Tom Hanks (i.e., a person), the target entity ID is Rita Wilson (i.e., a person), and the marriage has no end date because they are married after 1988 (i.e., a token). Thus, all constraints are satisfied and the sentence structure can be identified as valid. The voice output response could be "Tom Hanks is currently married to Rita Wilson since 1988".

Additional sentences can be added to the voice output response. Additional statements may be required based on the semantics of the query. For example, asking about Tom Hanks' wife to find out about the ex-wife can be useful to anyone. Additional triple sets may be associated with the speech input query. For example, an additional triple set can look like "Tom Hanks, Previous Marriage, Samantha Lewis". Additional sentences can be associated with the sentence structure. When identifying a potential sentence structure, the primary sentence structure (ie, "[source entity ID] is the [target entity ID] and current marital status since [token]]" can be identified so that the two sentences flow together have. The subsequent sentence structure may be "[Source Entity ID]" from [Token] to [Token] with [Target Entity ID]. " The constraints associated with this sentence structure may be as follows.

Constraint 1: Source Entity ID = Person

Constraint 2: Target Entity ID = Person

Constraint 3: Token = years

Constraint 4: The marriage end date between the source entity ID and the target entity ID

The additional constraint can be presented to further narrow the sentence structure such that the source entity ID is the same source entity ID as in the previous sentence and the target entity ID is different from the target entity ID in the previous sentence.

Each constraint in this example is satisfied because both Tom Hanks and Samantha Lewis are married and have a marriage end date. Thus, the voice output response could be "Tom Hanks was previously married to Samant Lewis from 1978 to 1987". In one embodiment, subsequent sentences may be associated with pronouns rather than re-describing the same entity. For example, "Tom Hanks has been married to Rita Wilson since 1988. Tom Hanks was previously married to Samantha Lewis from 1978 to 1987." He has since married Rita Wilson, who had previously married Samantha Lewis from 1978 to 1987. "

A further example may be a speech input query asking if X is a previous member of the Y group. The sentence structure associated with this speech input query may be "[Source Entity ID] is a previous member of [Target Entity ID] ". An exemplary constraint for this example may be:

Constraint 1: Source Entity ID = Person

Constraint 2: Target entity ID = party

Constraint 3: The source entity ID is no longer a member of the target entity ID.

The constraints ensure that the sentence structure is selected only as a valid sentence structure in the appropriate context. In the above example of asking for a previous member of a party, the sentence structure for a net worth (e.g., [source entity ID] net asset is [token]) is a constraint that the source is no longer a member of the target It will not be selected because condition 3 is not satisfied.

Referring now to FIG. 3, a flow diagram of an exemplary method 300 for generating a natural language output is shown. At block 302, a query is received. The query may be a speech input query. The query can also be a text input. At block 304, a response to the query is identified. This can be based on the meaning of the query. At block 306, the query and the response are mapped to a set of triples. The set of triples may represent the response in a machine-readable manner and be translated into natural language output. At block 308, the rules associated with the set of triples are identified. A rule may include one or more sentence structures and one or more constraints associated with each sentence structure. The associated one or more sentence structures and one or more constraints are identified at block 310. At block 312, it is determined if the constraints are satisfied. Each constraint associated with a sentence structure must be satisfied in order for the sentence structure to be a valid sentence structure. Based on the determination that the constraints are not satisfied, the method 300 returns to block 310 to identify other sentence structures and associated evaluated constraints. This process continues until the valid sentence structure is identified or the sentence structure of the evaluation subject is no longer present. If no valid sentence structure is identified, no output will be generated.

Based on the determination that the constraints are satisfied, the statement structure is identified as valid in block 314. At block 316, the output is passed in the form of a valid sentence structure. The output can be a single sentence or multiple sentences. The output can be a voice output so that the output speaks to the user.

Referring now to FIG. 4, a flow diagram of an exemplary method 400 for generating a natural language output is shown. At block 402, a query is received. The query may be a speech input query. At block 404, a response to the query is identified. At block 406, the response is mapped to the structured data. In one embodiment, the structured data is a collection of triples. At block 408, a sentence structure associated with the set of responses and triples is identified. At block 410, one or more constraints associated with the sentence structure are identified as satisfied. When one or more constraints are identified to be satisfied, the output response is conveyed in the form of a sentence at block 412.

Referring now to FIG. 5, a flow diagram of an exemplary method 500 for generating a natural language output is shown. At block 502, a speech input query is received. The voice query input may be interpreted or parsed for the knowledge graph path associated with the set of triples. At block 504, a response to the speech input query is identified. The response can be based on the interpreted knowledge graph path. At block 506, the response and the query are mapped to a set of triples. The set of triples can be associated with the knowledge graph path. At block 508, at least one rule associated with the set of triples is identified. The at least one rule includes one or more sentence structures associated with the set of triples and one or more constraints associated with the one or more sentence structures. In an embodiment, the sentence structure is a context free grammar sentence that satisfies all response triple constraints. At block 510, it is determined if at least one constraint is satisfied. When it is determined that at least one constraint is not satisfied, at block 514 the context free grammar sentence is identified as invalid and an additional sentence structure can be evaluated. Additional sentence structures may be evaluated until the valid sentence structure is identified or there is no longer any sentence structure identified for evaluation. Then, if the valid sentence structure is identified, the final sentence is constructed.

When it is determined that at least one constraint is satisfied, at block 512 the sentence structure is identified as a valid sentence structure. When a valid sentence structure is identified, at block 516, a speech output response is delivered.

While the invention has been described in connection with specific embodiments thereof, such embodiments are intended to be illustrative rather than limiting in all respects. Alternative embodiments will be apparent to those skilled in the art without departing from the scope of the present invention.

Claims (10)

In a method for configuring natural language output in a server,
Receiving a voice input query from a user;
Identifying a response to the speech input query;
Mapping the response to the structured data from a knowledge base arranged to store structured data, the mapping converting the response to a machine readable language used by a natural language engine; step;
Identifying a sentence structure associated with the structured data to which the response is mapped;
Arranging the structured data to which the response is mapped into the sentence structure;
Confirming that at least one constraint associated with the sentence structure is satisfied such that only a valid sentence structure is output;
Delivering an output response for the query in the form of a statement when each of the one or more constraints is met;
To construct a natural language output. The method according to claim 1,
Wherein the structured data is one or more triple sets. The method according to claim 1,
Wherein the output response is a one-sentence speech output along with a web search result page. The method of claim 3,
Wherein the output response further comprises a text output. A system for generating a natural language output in response to a query,
A computing device associated with a natural language engine having one or more processors and one or more computer storage media;
A memory including a data storage unit coupled with the natural language engine;
Lt; / RTI >
The natural language engine includes:
Identify the response to the query;
Translating the response to machine readable used by the natural language engine;
Map the response to the structured data from a knowledge base arranged to store the structured data;
Identify a sentence structure associated with the structured data;
Checking whether at least one constraint associated with the sentence structure is satisfied such that only a valid sentence structure is output;
Construct a sentence by arranging the structured data in a sentence structure associated with the structured data;
Wherein the output of the query is sent in the form of a sentence. 6. The method of claim 5,
Further comprising a query parser configured to identify the semantics of the query. 6. The method of claim 5,
Wherein the structured data is a set of triples. 6. The method of claim 5,
Wherein the natural language engine conveys an output response when one or more constraints associated with the sentence structure are satisfied. 6. The method of claim 5,
Wherein the natural language engine is configured to evaluate a plurality of sentence structures until the sentence structure with each constraint associated with the sentence structure is satisfied. A computerized method for generating natural language output in sentence form,
Receiving a query from a user, the query being a speech input query from the user;
Identifying a response to the speech input query;
Mapping the response to a set of triples, the mapping converting the response to a machine-readable language used by a natural language engine;
Identifying at least one rule associated with the set of triples, wherein the at least one rule comprises at least one constraint associated with a context free grammar sentence structure associated with a set of triples and the context free grammar sentence structure, Identifying the at least one rule;
Constructing an initial sentence by arranging the set of triples in a sentence-free grammar sentence associated with the set of triples;
Determining if at least one constraint associated with the context free grammar sentence structure is satisfied; And
Delivering a speech output response to the speech input query to a final sentence when determining that at least one constraint associated with the context free grammar sentence structure is satisfied
A natural language output generating computerized method, including:

Images